it’s time for a big idea · 2014-07-18 · new taus investments 1. free membership for...

It’s Time for a Big IdeaIt’s time for the Human Language Project

www.taus.net

We live in a time of hyper-globalization

Trade integration:

Last 2 decades 75 of developing countries are catching up with

economic frontier

World trade is growing

From stuff to fluff

Democratic: openness is widely embraced

Criss-crossing globalization

Peterson Institute for International Economics, July 2013

Global flow will double or triple before 2025

“Global flows are creating new degrees of connectedness among economies—and playing an ever-larger role in determining the fate of nations, companies, and individuals; to be unconnected is to fall behind. Our scenarios show that global flows could reach $54 trillion to $85 trillion by 2025, more than double or triple their current scale.”

McKinsey, April 2014

The History of Translation

The History of Translation

Print Democratized Knowlege

Computers and Internet Democratized Globalization

One Big Idea: Esperanto

Linguistic Diversity is the new Reality

This slide may not be used or copied without permission from TAUS

Another Big Idea....

Evolution of the Translation Industry

Big Bang of Content and Communications

Evolution of the Translation Industry

Translation Shifts Gears

21st Century Convergence

Luxury

Publisher-driventranslation industry

Mobile

Real-time

Personalized

Datafied Embedded

New payment models

Good enough

Continuous

+ 1,000 languages

Innovation Invaders

From 10,000 customers who buy translation as a ‘luxury’ product to 6 billion users who consider translation ‘free’.

Is the Translation Industry Ready?

Together

We Know

More

We Know

Better

Who do we (translation operators) work for?

For companies and governments?

.... or do we work for Global Citizens?

TAUS MEETS “GLOBAL CITIZEN” TARS

Moscow, May 24, 2014Tars speaks Bashkir & Russian

Drives a taxi in Moscow

Plays the qurai in his free time

Relies on Yandex for his translations

Will Tars find the time and ‘go the extra mile’ to learn English?

.... or will he continue to rely on Yandex to communicate with his fellow global citizen?

English or Machine Translation as the New Lingua Franca

More and more people will learn to speak English and

English will become the lingua franca of the world.43,3%

Linguistic diversity is the new reality and MT technology

will become the lingua franca of the world56,7%

Hello World! How do we communicate?

It’s Time for another.... Big Idea!

Human Language

ProjectCollaboration between business,

government and academia

worldwide

Think of: the Human Genome Project: a $3.8

Billion investment in sharing data

about the human genome drove

$796 Billion in economic impact,

and spurred growth in the life

sciences industry.

Translation is

Becoming a

Utility

MT as the New

Lingua Franca

Ingredients of Human Language Project

The Human Language Project consists of (at least):

1. Fearless sharing of language and translation data (speech and text) in all languages and

language pairs, not hindered by outdated copyright law. European legislators must modernize

copyright regulations on translation data. (See TAUS article published in January 2013)

2. A library of translation, language and reordering models covering all languages and a wide

scope of domains to help fast-track and fine-tune the development and customization of machine

translation engines.

3. A translation quality evaluation platform to help assess, benchmark and predict the right

translation quality for different content types and different purposes of communication.

4. A library of language tools – such as parsers, chunkers, lemmatizers, taggers – to assist service

and technology providers to improve and customize their solutions.

5. Common translation web services API’s to ensure that all services and technologies work

seamlessly together.

Building Blocks

This slide may not be used or copied without permission from TAUS

Human Language ProjectCollaboration between business,

government and academia worldwide

DataTranslation data of 400 most

spoken languages in the world

(160,000 language pairs).

TechnologyCentral resource for education

and advancement in open-

source and commercial MT.

APICoordinate translation web

services to optimize

connectivity.

EvaluationProvide for measuring and

benchmarking translation

quality.

Our aspiration is to enable better translation, to help the world

communicate better.

Our business is to support buyers and providers of translation with industry-

shared resources and technology.

TAUS Objectives

DataThe largest publicly shared

repository of translation data.

Currently 55 billion words.

TechnologyCentral resource for education

and advancement in

translation technology.

EvaluatePlatform for measuring and

benchmarking of translation

quality.

APICoordinate translation web

services to optimize

connectivity.

Four Action Lines

TAUS Service Overview

Search

Use scenarios

TAUS Data API

Search TMs

List software and MT

engines

Find tools and MT

engines

Find post-editors

Find specialists

Get data

Translate

Use scenarios

Knowledge Base

MT Users Call

Translate Community

Question-answering

Training

Translation API

Evaluate

Use scenarios

Knowledge Base

Content profiling

Evaluate Community

Question-answering

Professional evaluators

Training

Post-edit

Use scenarios

Knowledge Base

Post-editing Community

Question-answering

Professional post-editors

Training

New TAUS Investments

1. Free membership for universities

Use of 40 billion words in 2,000 language pairs for research only

Use of DQF & metrics for education

2. Translation QE standardization

Dynamic Quality Framework: shared industry platform

Work with ISO to convert best practices into standards for QE

3. Translation API standardization

Work with industry to enhance and standardize Translation API

4. New features for Data platform

Data Collector

Term Extractor

Matching Scores

Returns on Investment

1. Reduce the shortage of skills and talents

Free university program prepares new generation of language workers

2. Bring credibility, reliability and flexibility to industry

QE standard and industry benchmarking promotes scalability

3. Stimulate growth and innovation

Common translation API reduces friction and stimulates innovation

4. Proliferate use and development of MT

More shared data improves the performance of MT

Benefits both for individual companies and for the industry as a whole.

TAUS Events in Vancouver

TAUS Annual Conference

Vancouver, BC (Canada)

October 27-28

TAUS Quality Evaluation Summit

Vancouver, BC (Canada)

October 29

TAUS Moses Industry Roundtable

Vancouver, BC (Canada)

October 26 Afternoon

TAUS Machine Translation Showcase

Vancouver, BC (Canada)

October 29

TAUS Webinars

Translation Quality

Webinar

Bimonthly

Post-editing Webinar

WeeklyTranslation Technology

Showcase Webinar

Monthly

DQF Users call

Bimonthly

Translation Automation

Users Call

Monthly

This slide may not be used or copied without permission from TAUS